Adapting Listening Comprehension Materials through Generative Artificial Intelligence

Abstract

Listening comprehension has traditionally been one of the least attended skills in the teaching of Spanish as a foreign language, partly due to the gap between highly controlled but weakly authentic textbook materials and non-adapted authentic materials, which often

generate excessive cognitive load. In this context, the present study aimed to analyse the use of Generative Artificial Intelligence tools for the creation of listening materials adapted to the needs of learners of Spanish as a foreign language. Adopting an applied qualitative approach, the study described a systematic procedure for the selection of authentic source materials, their linguistic adaptation, and the subsequent generation of audio resources through large language models and text-to-speech tools. The analysis focused on controlling key variables for listening comprehension —such as lexis, syntactic complexity, speech rate, and prosodic clarity— in order to facilitate comprehensible input close to the i+1 level. The results showed that the generated materials reduced the cognitive load associated with non-adapted authentic audio while maintaining a functional degree of authenticity and curricular coherence. It was concluded that Generative Artificial Intelligence can constitute an effective complementary strategy for teaching listening comprehension in Spanish as a foreign language, provided that its use is critically integrated, guided by explicit pedagogical criteria, and mediated by teachers’ professional competence.